Same workspace
Humans and AI write to the same place. No parallel AI tree.
For people who live in Claude, Cursor, and ChatGPT. One shared memory across all of it. So no one re-explains context.
Every Claude or Cursor session starts blank. You re-paste the same stack, decisions, context. iri keeps it in one file. Every AI client reads it over MCP.
Everyone writes to it, humans and AI. A living, append-only file of every doc, decision, and conversation.
iri lives in Slack, Discord, Teams, and Google Chat.
No new app to learn.
A solo dev with three Claude tabs. A four-person team on Discord.
Same pattern: humans and AI working on the same things.
iri Notes is the workspace built for both.
Humans and AI write to the same place. No parallel AI tree.
Every client (Claude, Cursor, agents) reads it over MCP. Zero re-pasting.
Notes, decisions, threads in one file. Anyone reads on demand, human or AI. New teammates and agents start in-context.
Humans and AI write into the same memory and quote the same sources back. Claude reads Maya's Slack thread; Maya reads what Claude decided last night. One file, both sides, every claim traceable.
Each claim gains weight as new sources confirm it. Older decisions get superseded, not silently overwritten.
Ask once. iri returns the short, current answer. The length a teammate would write in Slack.
Threads, channels, AI sessions, docs flow in. iri organizes. The next teammate (human or AI) picks up where the last left off.
Slack / Discord threads sync in. Claude / Cursor / ChatGPT write back over MCP. Drag-and-drop docs auto-convert to AI-readable .md.
Notes are linked, deduped, and served as shared context. Available over MCP to every AI client.
The next teammate, human or AI, starts with that context already loaded. No re-pasting.
Notes, threads, code, decisions. All become atoms in one shared layer. Then every agent you work with reads from it.
For solo builders, small teams, and the AI agents you ship with. Free to start, no credit card.